DESCRIPTION (Applicant's abstract): Whereas affect has traditionally been conceptualized on a bipolar continuum ranging from positive to negative, Cacioppo and Berntson's (1994) bivariate model of evaluative space posits separable dimensions of positivity and negativity. The nature of evaluative processes bears upon our understanding of affective disorders, particularly in light of the notion that a general down- regulation of affect underlies depression. In bivariate terms such down-regulation may in fact be mediated by specific dysfunction of positive and/or negative activation. Our aim is to test whether the late positive potential (LPP) of the ERP can index the separability of positivity and negativity. Participants will make evaluative categorizations of very positive, mildly positive, and mildly negative trait words embedded in a mildly positive context. Insofar as the LPP indexes the separability of positivity and negativity, mildly negative (opposite-valence) targets will produce larger LPPs than very positive (same-valence) targets. Other participants will make evaluative categorizations of mildly positive, mildly negative, and very negative trait words embedded in a mildly negative context. Here mildly positive (opposite-valence) targets are expected to evoke larger LPPs than very negative (same-valence) targets. In both conditions, bipolar models predict that same-valence and opposite-valence targets will evoke comparably large LPPs.